Skip to main content
9 – 17 UHR +49 8031 3508270 LUITPOLDSTR. 9, 83022 ROSENHEIM
DE / EN
SPECIALIZED Google USA

Google MedGemma

MedGemma 1.5 by Google (January 2026) is an open-source AI for medical image and text analysis with 87.7% MedQA accuracy. EU-available via Vertex AI.

License Apache 2.0 (with usage restrictions)
GDPR Hosting Available
Modality Text, Image → Text

Versions

Overview of available model variants

ModelReleaseEUStrengthsWeaknessesStatus
MedGemma 1.5 27B Recommended
January 2026
87.7% MedQA Accuracy 3D medical imaging (CT/MRI) Longitudinal image analysis Comprehensive medical document processing
Not a certified medical device Higher computational requirements
Current
MedGemma 1.5 4B
January 2026
Compute-efficient, offline-capable 64.4% MedQA Accuracy Multimodal (text + image) Best performance in <8B category
Not a certified medical device Lower accuracy than 27B variant
Current

Use Cases

Typical applications for this model

Medical Research
Clinical Decision Support
Medical Image Analysis
Literature Review
Medical Education
Diagnostic Support

Technical Details

API, features and capabilities

API & Availability
Availability Public (Open Source)
Features & Capabilities
Structured Output Vision Reasoning Mode
Training & Knowledge
Knowledge Cutoff Mid-2025
Fine-Tuning Available (LoRA, Full Fine-Tuning, QLoRA)
Language Support
Best Quality English
Supported Primarily English, limited multilingual support
Medical terminology primarily in English

Hosting & Compliance

GDPR-compliant hosting options and licensing

GDPR-Compliant Hosting Options
Self-Hosted
Own Infrastructure
Recommended for patient data
Google Cloud
Frankfurt (europe-west3)
Vertex AI with EU data residency
License & Hosting
License Apache 2.0 (with usage restrictions)
Security Filters None (self-hosted responsibility)
On-Premise Edge-capable

Benchmarks

Performance comparison with standardized tests

MedQA
87.7%
MedQA
64.4%
DermMCQA
0.72
PathMCQA
0.70
Lab Report Extraction (Macro F1)
0.78
CT Scan Findings Accuracy
61%
MRI Findings Accuracy
65%

innFactory AI Consulting from Rosenheim, Germany advises enterprises across the DACH region (Germany, Austria, Switzerland) on GDPR-compliant integration of medical AI. MedGemma 1.5, released by Google in January 2026, is one of the most capable open-source models for healthcare applications - we strongly recommend self-hosting for patient data and careful compliance review under MDR/IVDR and the EU AI Act.

Important Notice

MedGemma is NOT an approved medical device. It must not be used as the sole basis for clinical decisions. All results must be validated by qualified medical professionals.

New Capabilities in MedGemma 1.5

Enhanced Imaging Support

  • 3D medical imaging: Processing complete CT and MRI volumes, not just individual slices
  • Whole-slide pathology: Analysis of large-scale histopathology images through patch-based processing
  • Longitudinal image analysis: Comparison of image series over time, such as multiple chest X-rays of the same patient
  • Improved anatomical localisation: Significantly more precise identification of anatomical structures, particularly in chest radiographs

Substantial Performance Improvements

MedGemma 1.5 demonstrates significant advances over previous versions:

  • CT findings: Accuracy increased from 58% to 61%
  • MRI findings: Accuracy improved from 51% to 65%
  • Laboratory reports: Macro F1 score for structured data extraction improved from 60% to 78%
  • Histopathology: ROUGE-L score increased from 0.02 to 0.49
  • Radiograph localisation: Intersection-over-Union improved from 3% to 38%

Two Optimised Variants

  • MedGemma 1.5 27B: For complex, text-heavy use cases with 87.7% MedQA accuracy
  • MedGemma 1.5 4B: Compute-efficient multimodal variant that can operate offline, with 64.4% MedQA accuracy

Availability and Integration

Cloud Deployment

MedGemma 1.5 is available through multiple channels:

  • Hugging Face: Direct downloads of open-source models
  • Google Vertex AI: Cloud deployment in EU regions (europe-west3 Frankfurt)
  • Model Garden: Integration into existing Google Cloud workflows

Complementary Technology: MedASR

Alongside MedGemma 1.5, Google released MedASR, a specialised speech-to-text model for medical dictation. This combination enables efficient clinical documentation with downstream AI analysis.

Regulatory Aspects

EU Regulation for Healthcare AI

  • MDR/IVDR: Medical device regulations must be strictly observed
  • EU AI Act: Medical AI classified as high-risk system
  • Liability: Full responsibility lies with the user, not the model developer
  • Validation: Clinical validation required before production deployment

Data Privacy in Healthcare Context

  • GDPR Art. 9: Special categories of personal data (health data)
  • Self-hosting recommended: Patient data should not leave your infrastructure
  • Anonymisation mandatory: Required before any processing with AI models
  • Data residency: When using cloud services, select EU regions

Self-Hosting for Maximum Control

For medical applications involving patient data, we expressly recommend self-hosting:

  • Complete data control and sovereignty
  • No data transmission to external providers
  • Fulfilment of stringent data protection requirements
  • Independence from cloud provider availability

The open-source nature of MedGemma enables flexible deployment in your own infrastructure, from on-premise servers to private cloud environments.

Our Recommendation

MedGemma 1.5 27B is suitable for research institutions and organisations with sufficient computational resources requiring highest accuracy. The 4B variant is ideal for resource-constrained environments or edge deployment.

Both variants are valuable tools for medical research, clinical education, and as support systems. However, they must never serve as the sole basis for clinical decisions.

For healthcare organisations, we offer specialised AI consulting - particularly regarding compliant integration under MDR, IVDR, and the EU AI Act.

Consultation for this model?

We help you select and integrate the right AI model for your use case.